Determining if pretreatment PSA doubling time predicts PSA trajectories after radiation therapy for localized prostate cancer.

Abstract

To determine if pretreatment PSA doubling time (PSA-DT) can predict post-radiation therapy (RT) PSA trajectories for localized prostate cancer. Three hundred and seventy-five prostate cancer patients treated with external beam RT without androgen deprivation therapy (ADT) were identified with an adequate number of PSA values. We utilized a linear mixed model (LMM) analysis to model longitudinal PSA data sets after definitive treatment. Post-treatment PSA trajectories were allowed to depend on the pre-RT PSA-DT, pre-RT PSA (iPSA), Gleason score (GS), and T-stage. Pre-RT PSA-DT had a borderline impact on predicting the rate of PSA rise after nadir (p=0.08). For a typical low risk patient (T1, GS6, iPSA 10), the predicted PSA-DT post-nadir was 21% shorter for pre-RT PSA-DT<24month compared to pre-RT PSA-DT>24month (19month vs. 24month). Additional significant predictors of post-RT PSA rate of rise included GS (p<0.0001), iPSA (p<0.0001), and T-stage (p=0.02). We observed a trend between rapidly rising pre-RT PSA and the post-RT post-nadir PSA rise. This effect appeared to be independent of iPSA, GS, or T-stage. The results presented suggest that pretreatment PSA-DT may help predict post-RT PSA trajectories.